duckdb/pg_duckdb

pg_duckdb is a PostgreSQL extension that integrates DuckDB's columnar-vectorized analytics engine into PostgreSQL to enable high-performance analytics and data-intensive applications.

The extension solves the problem of slow analytical queries on PostgreSQL by automatically routing them to DuckDB's optimized SQL engine when enabled. It requires no special syntax or data export: you write standard SQL against your existing PostgreSQL tables, set the configuration flag duckdb.force_execution=true, and the extension transparently executes the query using DuckDB's vectorized engine. Beyond querying PostgreSQL tables directly, the tool can read and write Parquet, CSV, JSON, Iceberg, and Delta Lake formats from cloud storage including S3, GCS, Azure, and R2, and supports joining PostgreSQL data with external data lake sources.

Adoption suits teams running analytical workloads on PostgreSQL who want performance improvements without migrating data or rewriting queries. It is particularly valuable for applications that combine transactional PostgreSQL tables with analytical queries, or for teams that need to query data lakes alongside their primary database. The extension includes optional integration with MotherDuck, a cloud analytics platform, allowing queries to execute on remote compute infrastructure while still accessing PostgreSQL tables.

The project maintains active development with regular releases and Docker image distributions. Development activity shows consistent engagement with the codebase through ongoing updates and refinements to the extension's capabilities.